>> In anybody figure out (and
>> comfirm) what NetBSD/pmax ULTRIX emulation can do or can't do?
>
> In the past I've run framemaker, but haven't tried Ultrix compat for a
> very long time.

I routinely use /usr/lib/cmplrs/cc3.0/dis to digout kernel binaries.
(sometimes, objdump).  When I tried it extensively (many years ago),
most of user applications could run well, including locally compiled
free-softwares, except for filesystem/device driver specific stuff.

> Now that hpcmips is here (I want one of these!)

When I can buy an execess, I will sent it to you.

> it may
> be good to move all the Ultrix compat goo to sys/arch/mips.  Should it
> be limited to little-endian platforms?  I know that most R4k onwards
> should be able to choose endian on a per-process level but don't know if
> we have plans to support this or how hard this would be...

R3000A processor seems to do runtime endian-switching as well.  Some of
Japaneses once toyed an idea of 'endian reversing emul layer', but it
didn't be realized.

>> It'd
>> also helpful to have 'howtos' to prepare /emul/ultrix environment
>> refering to the existense of /usr/share/sample/emul/ultrix/svc.conf.
>> How about emul.ultrix(5)?
>
> Umm, compat_ultrix(8) :-)

Arguable point.  Section 7 might be a candidate because a historic
heir(7) resides there.

Tohru Nishimura